On 2/21/19 6:22 AM, Sven Schnelle wrote: > Hi Richard, > > On Wed, Feb 20, 2019 at 10:57:50AM -0800, Richard Henderson wrote: > >> Can you re-test with the following? >> >> If the stores are all aligned, then the reads to not need to re-align. >> Also, partial writes to the config register don't make much sense. >> Over in hw/pci/pci_host.c, we check addr == 0 && len == 4, which is a >> very old fashioned way to do the .valid thing. > > With your proposed change it works as well - should i send a v2 or will you > take care of updating the patch?
I can squash this change. Thanks for testing. r~
